Github user smurching commented on the issue: https://github.com/apache/spark/pull/19106 @sethah I haven't heard of anybody hitting this issue in practice, but it did seem best to ensure that valid probability distributions would be produced regardless of input. There was some discussion of this in the JIRA: https://issues.apache.org/jira/browse/SPARK-21770
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org